OOPS! 3 Expensive Out-of-Stock Mistakes and How to Fix Them
80% of your product makes up only 20% of sales. Almost 30% of your product sells less than 1 per period. Managing the inventory for the 80% product mix is critical to raising profits and beating the competition. Retailers that continue to use top down replenishment strategies will not survive the new marketplace with the hyper-informed customer. Below are three ideas you can review in your business to help determine your opportunities for improvement with slow and intermittent demand products.
OOPS! 3 Expensive Out-of-Stock Mistakes and How to Fix Them
The Truth about Forecasting Slow Moving Products
Most people think they need multiple forecast algorithms or methods for managing slow and intermittent demand. In reality, there are two forecast methods that work very effectively: Croston’s and Poisson. A third method called ‘bootstrapping’ is being discussed in more circles as a possibility. Out-of-Stocks and overstocks occur more often due to poor practices by the retailer, not the forecast algorithm used. The origin for problems with any forecast method are very basic.
- Incorrect use of the forecast method
- Lack of accounting for seasonality
Sounds simple right? Here is the catch: many people don’t do either of these steps well, and the results are always out-of-stocks and overstocks, often in the same business lines at the same time. Bad data and mis-use of a formula can take on many variations.
Below are three ideas you can review in your business now. Three mistakes to avoid are:
- Re-forecast too often: use monthly, not weekly, when products have average or low demand.
- Failure to identify when a product has seasonal or market trends occurring.
- Wrong data used for forecasting regular demand
Re Forecast Frequency: Think about a slow mover that sells 1 unit every 4 weeks. If you run a forecast accuracy on 3 of the weeks, then the forecast is 100% wrong. This simple example is why forecasting across a broader period (4, 8, or 13 weeks) is very useful for slow moving products. A good forecast engine will use the zeros in the forecast math. If you measure the forecast weekly, then the system will look each week and say the forecast is 100% wrong in 3 of the 4 weeks of our example. That math will translate into more safety stock and greater error margins.
Seasonal & Market Trends: Often a product has seasonal or market trends that the human eye fails to catch. A index or trend is actually just a multiplier to the base forecast. The best way to determine an index is to use three years of regular sales data. Remember regular sales data must be devoid of promotions, events, and closeout sales. The sales should include regular sales and lost sales. Graph the three years by week and see if a pattern exits. If a pattern exists build the index (call us for help) and apply the index to your base forecast. This will flatten out what appears to be slow or intermittent demand in many cases.
Review your slow and intermittent demand products and determine if you have the right service goal, accurate forecast, and best re-forecast setting. If you need help, then call us and we are glad to step through some of your examples to help you learn best practices.
Wrong Data for Replenishment Creates Out-of-Stocks Forecast by type of demand: Break your sales into types and forecast each type of demand. For example, sales could rise rapidly (intermittent demand) but were due to a promotion. A sales forecasting system will raise the forecast, while a demand forecasting system will not because the regular type sales were the same; the promotion accounted for the sales rise.
80% of replenishment systems today use a sales forecast methodology, and, yes, some call themselves demand forecasting. (see our blog Differences in Demand and Sales Forecasting methods).
Imagine you have a product that regularly sells one a month. At the end of the month, you have a BOGO sale resulting in 4 sales for the month. This seemingly intermittent demand spike causes the sales forecast engine to raise the forecast and the safety stock. Reviewing sales next month and next year, the total sales is not a representation of regular demand. The total sales are representative of regular demand PLUS promotional demand. This gets more complicated when you include closeout sales, seasonal sales patterns, and lost sales. Demand Forecasting by sales type (regular, lost, promo, closeout) will flatten out many intermittent demand products.
Most buyers will fail to adjust the root cause, a bad forecast. Buyers will rely of their instincts and change a PO which results in overstocks or out-of-stocks. A human cannot effectively track the hundred or more products they manage through the supply chain with reports.
Demand Forecasting will deliver a more accurate forecast than a sales forecasting method due to better use of the data. The result is fewer out-of-stocks.
Copyright © Data Profits, Inc. 2013 All Rights Reserved.
Does 90% Forecasting Accuracy Sound Interesting to You?
Download our FREE white paper PDF to learn about true demand forecasting.
Leadership in Retail, Wholesale, and Software
“Data Profits iKIS was developed after years of inventory consulting with over 200 customers. Working with C level executives at many retail and wholesale establishments, we developed our unique and highly configurable BI dashboard, collaboration, and analysis software platform which provides demand forecasting, replenishment, lead time forecasting, optimization, and order management.”
Stuart Dunkin
Stuart graduated from Auburn University, with bachelor of science in business administration.
- Demand Forecasting: The Ultimate Secret for Your Organization’s Success - August 7, 2024
- How to Avoid Carrying Cost Mistakes in Inventory Optimization - June 10, 2024
- 3 Common Forecasting Software Issues and How to Fix - May 20, 2024
